November 27, 2004

The engineering in science, the experimentation in philosophy and art

Writing about an entry you don't have permission to view

"Experimental method does sometimes manipulate things" – no, it always manipulates things. Tell me of an experiment that doesn't involve the manipulation of a variable and the recording of the effect of that manipulation. Even in the case of experiments that seek to be purely the observation of 'natural events', those events are manipulated such that variables may be isolated and their relationships, as events occur, quantified. So every form of experiment is engineered, constructed, manufactured. And for every experiment, there is possibly an alternative configuration, a different engineering solution.

"Some of the most famous experiments in modern physics are gedankenexperiment – thought experiments – that don't actually require any prodding of stuff at all." – it depends on what you think that stuff is. A thought experiment considers a possible world in which variable A is somehow related to variable B. Some mechanism is posited that explains how the variables interact. Again the whole thing is engineered, this time as a simulated set of variables and functives. However, those simulated objects must still be constituted as mathematical, computational or conceptual objects so that they may be manipulated. For the thought experiment to have any 'scientific validity' that engineering must hold up to scrutiny, must work, in the same way as any engineered structure must work.

And here's the final twist: are physical experiments and scientific thought experiments different in kind, or only by degree? In the former case, a model is applied to the physical world in order to see if something is missing or inaccurate. The model predicts what will happen. Even when we are just seeking to gather empirical data, we apply a model, assuming that the things that we are measuring are the things that we think them to be and not some figment of our imagination (consider if the sense data that we are tracking turns out not to belong to the object that we assume it to be part of, and in fact is just there by coincidence). In the case of thought experiments, we are again looking to see if the relationships that we posit can be seen to necessarily result in the effect that we posit as the result of the interactions – or is there something missing from our model? Perhaps the only difference between scientific thought experiments, which I am calling simulations, and physical experiments, is that the former deals with a restricted and less complex environment.

Notice the shift of terminology there, from talking of scientific thought experiments to referring to them as simulations. At the moment i'm reading the section of What Is Philosophy? (Deleuze and Guattari) that deals with the difference between science and philosophy. Science we are told is concerned with functions and variable. Philosophy is the realm of concepts, which are quite different. I'm still working on this, but it seems that the key difference is that functions and variables represent reversibles, and concepts are irreversible. Anyhow, D&G seem to reserve the term 'thought experiment' for a mode of experimentation that works with concepts, and is therefore the domain of philosophy. This implies that they see a commonality between experimentation in science (be it physical or simulated) and experimentation in philosophy (with concepts):

To be sure, there is as much experimentation in the form of thought experiment in philosophy as there is in science, and being close to chaos, the experience can be overwhelming in both. What is Philosophy?, Deleuze and Guattari 1994, p. 127

The different modes of experimentation are defined by their relationship with chaos. In the case of philosophy, the aim is to pass through chaos such that an impossible world is actualised, a formerly inconcievable state is reached. In the case of science, experimentation posits a set of possible states that can be interchanged for each other through the operation of a function along variables. The model opens itself to chaos through its application, and on the discovery of missing elements, assimilates them back into the model as further possibilities, as further variables. In this way science progresses, whereas philosophy differentiates.

But in reality science and philosophy are mixed. Scientific models can suddenly breakdown upon their engagement with chaos. And creation (or differentiation, its pseudonymn in D&G) then occurs:

But there is also as much creation in science as there is in philosophy or the arts. ibid p.127

The tendency of science, it could be said, is to sometimes creatively fictionalise the sense of progression by reinventing its model in a new but familiar form. Philosophy, of course, also suffers from this delusion, and must itself be more prepared to abandon the concept of progress. Art, the third of the trinity of disciplines, suffers from a related delusion, in this case the idea that it proceeds without experimentation, simply through a sensus communis of good taste experienced as genius:

There is no creation without experiment. ibid p.127

- 3 comments by 1 or more people Not publicly viewable

  1. A non-philosopher comments.
    Scientific experiment uses control experiments (of which the placebo is an example). They can (indeed must) also be repeated and a range of results recorded; and be repeatable by individuals with no knowledge of the process (or theory) underlying them, and provide the same results. Is this very empirical set of observations compatible with what you say in the posting?

    27 Nov 2004, 23:49

  2. I think it is. Surely even a control 'experiment' must be open to failure, contamintion or such, the whims and vagaries of chance.

    28 Nov 2004, 01:30

  3. All this sounds very interesting to me. I am trying to relate this with other philosophers (I always like to compare something new I read with what I have already read, so that I can situate it in my own thinking) and at a first glance I seem to see a similarity with Wittgenstein in the sense that Scientific inquiry, as well as philosophical thought experiments, is like a game with pre-established (presupposed) rules.
    There is only one kind of winner: the one who qualifies for the pre-set criteria. In footbal, thw winner is never going to be which team has the cleverest players, but the ones who actually scored most goals: the referee counts goals, not IQ ratings.

    The question, that is, the form of the question, essentially opens the way for the answer to develop. In effect, we should be busying oursleves with formulating the right questions, and re-evaluating our existing questions, instead of stambling on the answers.

    I'll comment more when I am sober, it's Saturday night 2:00 in the morning, and I don't have a clue if what I am saying is non-sense. Cheers.

    28 Nov 2004, 02:15

Add a comment

You are not allowed to comment on this entry as it has restricted commenting permissions.